English español
Búsqueda
 

Repositorio de la Universidad de Oviedo > Producción Bibliográfica de UniOvi: RECOPILA > Artículos >

Use este identificador para citar o enlazar este ítem: http://hdl.handle.net/10651/7187

Título : Enhancing directed binary trees for multi-class classification
Autor(es) y otros: Montañés Roces, Elena
Barranquero Tolosa, José
Díez Peláez, Jorge
Coz Velasco, Juan José del
Fecha de publicación : 2013
Editorial : Elsevier
Versión del editor: http://dx.doi.org/10.1016/j.ins.2012.10.011
Citación : Information Sciences, 223, p. 42-55 (2012); doi:10.1016/j.ins.2012.10.011
Descripción física: p. 42-55
Resumen : One approach to multi-class classi cation consists in decomposing the original problem into a collection of binary classi cation tasks. The outputs of these binary classi ers are combined to produce a single prediction. Winner-takesall, max-wins and tree voting schemes are the most popular methods for this purpose. However, tree schemes can deliver faster predictions because they need to evaluate less binary models. Despite previous conclusions reported in the literature, this paper shows that their performance depends on the organization of the tree scheme, i.e. the positions where each pairwise classi er is placed on the graph. Di erent metrics are studied for this purpose, proposing a new one that considers the precision and the complexity of each pairwise model, what makes the method to be classi er-dependent. The study is performed using Support Vector Machines (SVMs) as base classi ers, but it could be extended to other kind of binary classi ers. The proposed method, tested on benchmark data sets and on one real-world application, is able to improve the accuracy of other decomposition multi-class classi ers, producing even faster predictions. Keywords: Multi-class classi cation, Decomposition methods, Support Vector Machines, Directed Binary Trees, Generalization error bounds
URI : http://hdl.handle.net/10651/7187
ISSN : 0020-0255
Aparece en las colecciones: Artículos
Informática

Ficheros en este ítem:

Fichero Descripción Tamaño Formato
Enhancing Directed Binary Trees for Multi-class Classification.pdfPostprint509,78 kBAdobe PDFVisualizar/Abrir


Exportar a Mendeley


Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons
Creative Commons

Los ítems de DSpace están protegidos por copyright, con todos los derechos reservados, a menos que se indique lo contrario.

 

Base de Datos de Autoridades Biblioteca Universitaria Consultas / Sugerencias